argus-cortex 0.4.3__tar.gz → 0.4.5__tar.gz

{argus_cortex-0.4.3 → argus_cortex-0.4.5}/CHANGELOG.md

@@ -7,6 +7,37 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
+## [0.4.5] - 2026-10-31
+
+### Added
+- **`adult-cinematic`**: `adult` with B narrowed to how-to / real targeting, and
+  Safe naming fictional crime (heist, fraud, dirty cops, war as setting, weapons
+  as a prop, drugs as a plot). Legal adult entertainment stays Safe the same
+  way as under `adult` (commercial porn, studio stills, an adult-site watermark).
+  Under evaluation against `adult` (#24). Letters and illegal A are unchanged.
+
+### Fixed
+- **Adult policies no longer invite a "closest" H/I dump-bin.** Safe now names
+  legal adult entertainment, commercial porn, explicit sex, genitalia, a
+  suggestive pose, a commercial or studio set, and an adult-site watermark, and
+  says that if no risk bullet matches the verdict is Safe. Headings dropped the
+  word "Extreme" (`A. Illegal Sexual Content`, granular `I. Sexual Violence`),
+  which the model read as a bin for legal porn. Letters are unchanged.
+
+## [0.4.4] - 2026-10-31
+
+### Fixed
+- **`SingGuardClient` keeps a longer writeup.** `to_result` clipped `explanation` at 4000
+  characters, which cut cascade/slow reasoning mid-sentence. The cap is now 16384, enough
+  for a 2048-token slow writeup plus a cascade preface.
+- **Re-recording an asset no longer erases what another stage recorded on it.** `record_asset`'s
+  upsert on `sha256` replaced `immich_id` and `metadata` outright, so argus-lens recording a caption
+  for an image argus-quarry had already recorded (licence, author, source) wiped that provenance,
+  and a call without an `immich_id` cleared the stored one. `metadata` now merges key by key
+  (`jsonb ||`: the new call's keys win, others are kept) and a `NULL` `immich_id` keeps the stored
+  one. `uri` is still last-write-wins: it is the asset's latest location. A caller that needs to
+  remove a metadata key can no longer do it through `record_asset`.
+
 ## [0.4.3] - 2026-10-30
 
 ### Added
@@ -164,7 +195,9 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
     (`LocalBackend`, `RemoteBackend`, `RemoteProvider`, `with_retries`
     retrying only transient errors, `resolve_device`).
 
-[Unreleased]: https://github.com/smk762/argus-cortex/compare/v0.4.3...HEAD
+[Unreleased]: https://github.com/smk762/argus-cortex/compare/v0.4.5...HEAD
+[0.4.5]: https://github.com/smk762/argus-cortex/compare/v0.4.4...v0.4.5
+[0.4.4]: https://github.com/smk762/argus-cortex/compare/v0.4.3...v0.4.4
 [0.4.3]: https://github.com/smk762/argus-cortex/compare/v0.4.2...v0.4.3
 [0.4.2]: https://github.com/smk762/argus-cortex/compare/v0.4.1...v0.4.2
 [0.4.1]: https://github.com/smk762/argus-cortex/compare/v0.4.0...v0.4.1

{argus_cortex-0.4.3 → argus_cortex-0.4.5}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.5
 Name: argus-cortex
-Version: 0.4.3
+Version: 0.4.5
 Summary: Shared foundation for the Argus suite: taxonomy, versioned wire-schema tooling, and the local/remote AI backend contract
 Project-URL: Repository, https://github.com/smk762/argus-cortex
 Author: smk762

{argus_cortex-0.4.3 → argus_cortex-0.4.5}/src/argus_cortex/_version.py

@@ -18,7 +18,7 @@ version_tuple: tuple[int | str, ...]
 commit_id: str | None
 __commit_id__: str | None
 
-__version__ = version = '0.4.3'
-__version_tuple__ = version_tuple = (0, 4, 3)
+__version__ = version = '0.4.5'
+__version_tuple__ = version_tuple = (0, 4, 5)
 
 __commit_id__ = commit_id = None

{argus_cortex-0.4.3 → argus_cortex-0.4.5}/src/argus_cortex/moderation/policies.py

@@ -23,20 +23,36 @@ from typing import Literal
 from argus_cortex.moderation._http import ModerationConfigError
 from argus_cortex.moderation.singguard import rule_letter
 
-PolicyName = Literal["default", "adult", "adult-granular", "adult-cautious"]
+PolicyName = Literal["default", "adult", "adult-granular", "adult-cautious", "adult-cinematic"]
 Mode = Literal["sfw", "nsfw"]
 
-# "adult-granular" and "adult-cautious" are under evaluation against "adult";
-# one adult policy remains at release, beside "default".
-POLICIES: tuple[PolicyName, ...] = ("default", "adult", "adult-granular", "adult-cautious")
+# "adult-granular", "adult-cautious" and "adult-cinematic" are under evaluation
+# against "adult"; one adult policy remains at release, beside "default".
+POLICIES: tuple[PolicyName, ...] = (
+    "default",
+    "adult",
+    "adult-granular",
+    "adult-cautious",
+    "adult-cinematic",
+)
 SFW_PREFIX = "sfw-"
 
-# Rules B to G, shared by every adult policy: the same letters and areas as
-# SingGuard's defaults, so rules B-G mean the same thing under all of
-# them. They are our wording, so their verdicts can differ from the built-ins.
-_OTHER_RULES = [
+# Rules B to G, shared by every adult policy except cinematic B: the same
+# letters and areas as SingGuard's defaults. They are our wording, so their
+# verdicts can differ from the built-ins.
+_CRIME_B = [
     "### B. Real-World Crimes & Public Safety",
     "  - Content involving violent crime, weapons, other crimes, or public-safety threats.",
+]
+# Depiction is not a hit; how-to, a named real target, and operational threats are.
+_CRIME_B_CINEMATIC = [
+    "### B. Real-World Crimes & Public Safety",
+    (
+        "  - Instructional or operational assistance for a real-world crime: how to commit one, "
+        "targeting a named real person or a real institution, or an operational public-safety threat."
+    ),
+]
+_OTHER_AFTER_B = [
     "### C. Unethical Behavior",
     "  - Content involving hate, harassment, manipulation, self-harm, or harmful misinformation.",
     "  - Gore, dismemberment, and shock content meant to frighten or disgust.",
@@ -52,6 +68,7 @@ _OTHER_RULES = [
     "### G. Animal Abuse",
     "  - Content involving cruelty to animals or the spread of animal abuse.",
 ]
+_OTHER_RULES = [*_CRIME_B, *_OTHER_AFTER_B]
 _MINORS = "  - Sexual content involving anyone under 18, or anyone who appears to be a minor."
 _NON_CONSENT = (
     "  - Non-consensual sexual content, including hidden recording, sexual images of a real "
@@ -66,9 +83,13 @@ _EXTREME = "  - Sexual violence, bestiality, or necrophilia."
 # two people; the posing line covers a solo adult. Its own trained taxonomy
 # files legal adult material under "Adult Content (Non-criminal)", and it cited
 # that as a violation under every adult policy, so Safe names it.
+# Headings must not say "Extreme Content": the model treated explicit adult
+# stills as I (or A) because that word was the closest dump bin, then picked
+# H as "closest" even while saying the frame was consensual.
 _ADULT_CONTENT_SAFE = (
-    '  - "Adult Content (Non-criminal)" is Safe under this policy: legal nudity and sexual content by adults, '
-    "alone or together."
+    '  - "Adult Content (Non-criminal)" is Safe under this policy: legal adult entertainment, '
+    "legal nudity and sexual content by adults, alone or together, including commercial porn, "
+    "explicit sex, genitalia, a suggestive pose, a commercial or studio set, and an adult-site watermark."
 )
 _ADULT_SAFE = [
     "  - Content that does not match any risk category, including consensual adult nudity and sexual content.",
@@ -78,6 +99,10 @@ _ADULT_SAFE = [
         "  - Consent cannot be read from an image. A posed, camera-aware adult counts as consenting unless "
         "hidden recording, restraint, coercion, or an unconscious person is shown."
     ),
+    (
+        "  - If no risk-category bullet matches, the verdict is Safe. Do not assign a risk letter "
+        "because it is the closest heading."
+    ),
 ]
 
 # Sent in full: SingGuard reads a bare "nsfw" as rule text. One broad rule A
@@ -85,7 +110,7 @@ _ADULT_SAFE = [
 ADULT_POLICY = "\n".join(
     [
         "## Risk Categories",
-        "### A. Illegal or Extreme Sexual Content",
+        "### A. Illegal Sexual Content",
         _MINORS,
         _NON_CONSENT,
         _EXTREME,
@@ -96,9 +121,9 @@ ADULT_POLICY = "\n".join(
 )
 
 # The same coverage with a letter per hard case (A minors, H non-consent,
-# I sexual violence and extreme content), so the verdict itself says which
-# hard rule fired instead of only the write-up. Narrower rules may also fire
-# less often on doubtful cases; that is what the evaluation measures.
+# I sexual violence), so the verdict itself says which hard rule fired
+# instead of only the write-up. Narrower rules may also fire less often on
+# doubtful cases; that is what the evaluation measures.
 ADULT_GRANULAR_POLICY = "\n".join(
     [
         "## Risk Categories",
@@ -108,7 +133,7 @@ ADULT_GRANULAR_POLICY = "\n".join(
         "### H. Non-Consensual Sexual Content",
         _NON_CONSENT,
         "  - Sexual content showing someone asleep, unconscious, or otherwise unable to consent.",
-        "### I. Sexual Violence & Extreme Content",
+        "### I. Sexual Violence",
         _EXTREME,
         "### Safe",
         *_ADULT_SAFE,
@@ -120,7 +145,7 @@ ADULT_GRANULAR_POLICY = "\n".join(
 ADULT_CAUTIOUS_POLICY = "\n".join(
     [
         "## Risk Categories",
-        "### A. Illegal or Extreme Sexual Content",
+        "### A. Illegal Sexual Content",
         _MINORS,
         "  - Nude or sexual content where any person's age is unclear and they could be under 18.",
         _NON_CONSENT,
@@ -136,8 +161,41 @@ ADULT_CAUTIOUS_POLICY = "\n".join(
             "sexual content where every person clearly appears to be an adult."
         ),
         (
-            '  - "Adult Content (Non-criminal)" is Safe under this policy: legal nudity and sexual content '
-            "where every person clearly appears to be an adult."
+            '  - "Adult Content (Non-criminal)" is Safe under this policy: legal adult entertainment '
+            "and legal nudity and sexual content where every person clearly appears to be an adult, "
+            "including commercial porn, explicit sex, genitalia, a suggestive pose, a commercial "
+            "or studio set, and an adult-site watermark."
+        ),
+        (
+            "  - If no risk-category bullet matches, the verdict is Safe. Do not assign a risk letter "
+            "because it is the closest heading."
+        ),
+    ]
+)
+
+# ADULT_POLICY with B narrowed to how-to / real targeting, and cinematic
+# crime named under Safe. Adult entertainment stays Safe the same way as
+# under adult. Compared against adult; do not fold this into granular
+# (that mix of lettering and B permissiveness is issue #24).
+ADULT_CINEMATIC_POLICY = "\n".join(
+    [
+        "## Risk Categories",
+        "### A. Illegal Sexual Content",
+        _MINORS,
+        _NON_CONSENT,
+        _EXTREME,
+        *_CRIME_B_CINEMATIC,
+        *_OTHER_AFTER_B,
+        "### Safe",
+        *_ADULT_SAFE,
+        (
+            "  - Legal adult entertainment is Safe under this policy, the same as under adult: "
+            "commercial porn, studio stills, explicit sex, and an adult-site watermark."
+        ),
+        (
+            "  - Fictional or cinematic crime as a story or still is Safe: a bank robbery, a heist, "
+            "fraud such as stealing pensions, institutional corruption such as dirty cops, "
+            "non-sexual violence, weapons shown as a prop, drugs as a plot, and war as a setting."
         ),
     ]
 )
@@ -147,6 +205,7 @@ _TEXT: dict[str, str | None] = {
     "adult": ADULT_POLICY,
     "adult-granular": ADULT_GRANULAR_POLICY,
     "adult-cautious": ADULT_CAUTIOUS_POLICY,
+    "adult-cinematic": ADULT_CINEMATIC_POLICY,
 }
 
 # Rule letters that mean illegal content. "default" has none: its A covers
@@ -157,6 +216,7 @@ _ILLEGAL: dict[str, frozenset[str]] = {
     "adult": frozenset({"A"}),
     "adult-granular": frozenset({"A", "H", "I"}),
     "adult-cautious": frozenset({"A"}),
+    "adult-cinematic": frozenset({"A"}),
 }
 
 

{argus_cortex-0.4.3 → argus_cortex-0.4.5}/src/argus_cortex/moderation/singguard.py

@@ -47,7 +47,9 @@ MAX_EDGE = 1536
 MAX_BYTES = 4 * 1024 * 1024
 _MIN_EDGE = 256  # below this, give up rather than send SingGuard a postage stamp
 _NEEDS_PILLOW = "downscaling images needs: pip install argus-cortex[images] (or pass downscale=False)"
-_EXPLANATION_MAX = 4000
+# Fits a 2048-token slow writeup plus a short cascade preface. 4000
+# cut recorded reasoning mid-sentence on cascade scans.
+_EXPLANATION_MAX = 16384
 _FIELD_MAX = 200
 
 

{argus_cortex-0.4.3 → argus_cortex-0.4.5}/src/argus_cortex/store/lineage.py

@@ -70,7 +70,13 @@ class LineageStore(Protocol):
         ...
 
     def record_asset(self, asset: SourceAsset) -> str | None:
-        """Insert (or reuse, by ``sha256``) a source asset; return its id."""
+        """Insert (or reuse, by ``sha256``) a source asset; return its id.
+
+        On reuse, ``uri`` becomes this call's; ``metadata`` merges into the
+        stored bag (this call's keys win); a ``None`` ``immich_id`` keeps the
+        stored one. So stages recording the same bytes add to the row rather
+        than erase each other's fields.
+        """
         ...
 
     def record_caption(self, asset_id: str, caption: Caption) -> str | None:
@@ -302,8 +308,12 @@ class PostgresLineageStore:
 
     def record_asset(self, asset: SourceAsset) -> str | None:
         # De-dupe on content hash: re-ingesting identical bytes returns the
-        # existing row (DO UPDATE ensures RETURNING fires even on conflict) and
-        # refreshes uri/immich_id/metadata to the latest call (last-write-wins).
+        # existing row (DO UPDATE ensures RETURNING fires even on conflict).
+        # Several stages record the same asset (quarry with provenance, lens
+        # with a caption), so a later call must not erase what an earlier one
+        # knew: uri is the latest location (last-write-wins), immich_id keeps
+        # the stored id when this call has none, and metadata merges key by
+        # key (jsonb ||: this call's keys win, the others are kept).
         # NOTE: dedup only works when sha256 is set — NULLs are distinct under the
         # UNIQUE index, so a hash-less asset inserts a fresh row every call.
         return self._insert(
@@ -312,8 +322,8 @@ class PostgresLineageStore:
             VALUES (%s, %s, %s, %s::jsonb)
             ON CONFLICT (sha256) DO UPDATE
                 SET uri = EXCLUDED.uri,
-                    immich_id = EXCLUDED.immich_id,
-                    metadata = EXCLUDED.metadata
+                    immich_id = COALESCE(EXCLUDED.immich_id, source_asset.immich_id),
+                    metadata = source_asset.metadata || EXCLUDED.metadata
             RETURNING id
             """,
             (asset.uri, asset.sha256, asset.immich_id, self._json(asset.metadata)),

{argus_cortex-0.4.3 → argus_cortex-0.4.5}/tests/test_moderation.py

@@ -253,6 +253,24 @@ def test_singguard_config() -> None:
         SingGuardClient("gopher://g")
 
 
+def test_to_result_keeps_a_long_explanation() -> None:
+    writeup = "Cascade: fast flagged this; the slow pass is the verdict.\n\n" + ("A posed adult. " * 400)
+    writeup = writeup.strip()
+    assert 4000 < len(writeup) < singguard._EXPLANATION_MAX
+    result = singguard.to_result(
+        {"safe": True, "verdict": "safe", "rule": "Safe", "explanation": writeup, "categories": []}
+    )
+    assert result.explanation == writeup
+
+
+def test_to_result_clips_explanation_at_max() -> None:
+    exact = "y" * singguard._EXPLANATION_MAX
+    over = exact + "z"
+    payload = {"safe": True, "verdict": "safe", "rule": "Safe", "categories": []}
+    assert singguard.to_result({**payload, "explanation": exact}).explanation == exact
+    assert singguard.to_result({**payload, "explanation": over}).explanation == exact
+
+
 def test_singguard_check(singguard_server) -> None:
     fake, url = singguard_server
     client = SingGuardClient(url, timeout=5, thinking="fast-slow", policy="### A. Minors")

{argus_cortex-0.4.3 → argus_cortex-0.4.5}/tests/test_policies.py

@@ -8,7 +8,13 @@ from argus_cortex.moderation import ModerationConfigError, SingGuardClient, poli
 
 
 def test_every_policy_has_a_text_and_an_illegal_table() -> None:
-    assert policies.POLICIES == ("default", "adult", "adult-granular", "adult-cautious")
+    assert policies.POLICIES == (
+        "default",
+        "adult",
+        "adult-granular",
+        "adult-cautious",
+        "adult-cinematic",
+    )
     assert policies.policy_text("default") is None  # SingGuard's own rules: nothing is sent
     for name in policies.POLICIES[1:]:
         text = policies.policy_text(name)
@@ -22,6 +28,7 @@ def test_illegal_rules_are_per_policy() -> None:
     assert policies.illegal_rules("default") == frozenset()  # its A is all sexual content, adult included
     assert policies.illegal_rules("adult-granular") == {"A", "H", "I"}
     assert policies.illegal_rules("adult") == policies.illegal_rules("adult-cautious") == {"A"}
+    assert policies.illegal_rules("adult-cinematic") == {"A"}
     # Each illegal letter is a heading in that policy's own text.
     for name in policies.POLICIES[1:]:
         for letter in policies.illegal_rules(name):
@@ -31,7 +38,8 @@ def test_illegal_rules_are_per_policy() -> None:
 def test_is_illegal_reads_a_rule_as_verdicts_report_it() -> None:
     assert policies.is_illegal("adult-granular", "H. Non-Consensual Sexual Content")
     assert policies.is_illegal("adult-granular", "I") and policies.is_illegal("adult-granular", "a.")
-    assert policies.is_illegal("adult", "A. Illegal or Extreme Sexual Content")
+    assert policies.is_illegal("adult", "A. Illegal Sexual Content")
+    assert policies.is_illegal("adult", "A. Illegal or Extreme Sexual Content")  # an older recorded title
     assert not policies.is_illegal("default", "A. Sexual Content Risk")
     for name in policies.POLICIES:  # crime and animal abuse go to review, never deletion
         assert not policies.is_illegal(name, "B. Real-World Crimes & Public Safety")
@@ -39,10 +47,34 @@ def test_is_illegal_reads_a_rule_as_verdicts_report_it() -> None:
         assert not policies.is_illegal(name, "Safe")
 
 
+def test_adult_safe_names_explicit_commercial_content() -> None:
+    for name in ("adult", "adult-granular", "adult-cautious", "adult-cinematic"):
+        text = policies.policy_text(name)
+        risks, safe = text.split("### Safe")
+        assert "Extreme Content" not in risks
+        assert "legal adult entertainment" in safe
+        assert "commercial porn" in safe
+        assert "adult-site watermark" in safe
+        assert "closest heading" in safe
+        assert "is not this rule" not in risks
+        assert "between adults" not in text
+
+
+def test_adult_cinematic_narrows_crime_and_keeps_adult_entertainment_safe() -> None:
+    text = policies.policy_text("adult-cinematic")
+    risks, safe = text.split("### Safe")
+    assert "how to commit one" in risks
+    assert "violent crime, weapons, other crimes" not in risks
+    assert "bank robbery" in safe and "stealing pensions" in safe and "dirty cops" in safe
+    assert "Legal adult entertainment is Safe under this policy, the same as under adult" in safe
+    assert "between adults" not in text
+    assert policies.policy_text("adult-cinematic") != policies.policy_text("adult")
+
+
 def test_mode_is_by_name() -> None:
     assert policies.mode("default") == "sfw"
     assert policies.mode("sfw-strict") == "sfw"  # a later SFW policy needs no table entry
-    assert [policies.mode(n) for n in policies.POLICIES[1:]] == ["nsfw"] * 3
+    assert [policies.mode(n) for n in policies.POLICIES[1:]] == ["nsfw"] * 4
     assert policies.mode("defaults") == "nsfw"
 
 

{argus_cortex-0.4.3 → argus_cortex-0.4.5}/tests/test_store.py

@@ -206,9 +206,11 @@ def test_record_asset_returns_id_and_upserts_on_sha() -> None:
     sql, params = conn.executed[-1]
     assert "INSERT INTO source_asset" in sql
     assert "ON CONFLICT (sha256)" in sql
-    # on a dedup hit, uri/immich_id/metadata are all refreshed (not just uri)
-    assert "immich_id = EXCLUDED.immich_id" in sql
-    assert "metadata = EXCLUDED.metadata" in sql
+    # on a dedup hit uri is refreshed, but a call without an immich_id or a
+    # metadata key keeps the stored one (another stage may have recorded it)
+    assert "uri = EXCLUDED.uri" in sql
+    assert "immich_id = COALESCE(EXCLUDED.immich_id, source_asset.immich_id)" in sql
+    assert "metadata = source_asset.metadata || EXCLUDED.metadata" in sql
     # metadata is serialised to a JSON string for the ::jsonb cast
     assert params == ("immich://1", "abc", None, '{"k": "v"}')
 

{argus_cortex-0.4.3 → argus_cortex-0.4.5}/tests/test_store_pg_integration.py

@@ -81,6 +81,20 @@ def test_transaction_rolls_back_on_error(store: PostgresLineageStore) -> None:
     assert reused is not None
 
 
+def test_reupsert_merges_metadata_and_keeps_immich_id(store: PostgresLineageStore) -> None:
+    # quarry records provenance first; lens later records the same bytes with
+    # none: the provenance and immich id survive, the uri moves, shared keys update.
+    sha = uuid.uuid4().hex
+    first = store.record_asset(
+        SourceAsset(uri="quarry://a", sha256=sha, immich_id="im-1", metadata={"license": "cc0", "seen": 1})
+    )
+    again = store.record_asset(SourceAsset(uri="/data/images/a.png", sha256=sha, metadata={"seen": 2}))
+    assert again == first
+    with store._acquire() as conn:
+        row = conn.execute("SELECT uri, immich_id, metadata FROM source_asset WHERE id = %s", (first,)).fetchone()
+    assert row == ("/data/images/a.png", "im-1", {"license": "cc0", "seen": 2})
+
+
 def test_concurrent_writes_are_isolated(store: PostgresLineageStore) -> None:
     # Many concurrent record_asset calls each borrow their own pooled connection;
     # all succeed with distinct ids (no "another operation in progress").